like the error shows, the VM you are trying to provision fails to connect to mirror.centos.org.
This can have many reasons, most (if not all) probably related to your local network.
The things that come to my mind are:
Does your DHCP offer a valid DNS Server that can provide the address for mirror.centos.org?
Do you have a Firewall blocking access to the internet?
Are you using a proxy to access the internet?
These 3 are the most common causes of the error you are encountering. If I remember correctly, anaconda should drop you into an emergency shell after that error where you can poke around at the error.
Hi Areyus,
thank you for reply!
For the 1. and 2. question the settings are ok and firewall was temporary disabled.
For the 3. question I have modified the kickstart default with my proxy but network is always unreachable.
remote diagnosis of network related problems is kinda hard, but I will try to help.
First of, does dracut drop you into an emergency shell after it failed to download the image?
If so, the first steps you should take are:
Try if you can ping the gateway of your network
Try if you can ping your proxy
If any of those failes, you have a problem on your local network that needs to be fixed.
If those succeed, try to curl/wget (whatever is present in your installer image) anything from the internet (google, theforeman.org, whatever). If this failes, you likely have a problem with your proxy.
If you are not dropped to an emergency shell, you can try and boot the system from a Live ISO yourself and perform these steps from there.
After checking my network I confirm that I haven’t a proxy connection between foreman and provisioning VM. I also fill the proxy parameters in kickstart but not work.
How can I force the proxy?
fromthe kickstart documentation (Section B 2.21), you need to add the --proxy= parameter with your proxy’s connection details to the url command in your kickstart provisioning template.
From the error message you posted earlier, it looks like the connection to Foreman works fine (otherwise the VM could not have loaded the kickstart or the installer), but needs the proxy information to access the installer Image from the internet.
This is what should work in my opinion. Do you still get the same error when you inculde the proxy parameter in Kickstart default?
If you do not manage to get it to work with the upstream image, you always have the option to download the install image from centos and host it yourself on a local webserver. You will have to update that image manually in that case if it changes upstream, though (or write a script that automates this, obviously). If you want to do that, you need to change the URL of the medium in Foreman under “Hosts” → “Installation Media” → → Path to point to your local copy.
From what it looks like, I guess anaconda is trying to download the whole squashfs.img and runs out of space “somewhere” (maybe in RAM?)
How much RAM does you VM have? An installation of current CentOS/RHEL 7 requires the system to have somewhere around 1,5-2 GB of RAM. If you assigned less RAM, try setting it to at least 2 GB of RAM and try again.
Glad to hear you got it working
If this issue is solved, could you please mark the solution, so the thread is marked as “solved” an others know this is sorted out? There should be a checkbox with the text “solution” next to it to mark a post as solution.
